1 00:00:06,830 --> 00:00:12,410 ‫Well come to the microphone interface explain lecture in this lecture I'm going to explain the output 2 00:00:12,710 --> 00:00:17,210 ‫from a microphone and how we can read the signals with an FPGA. 3 00:00:17,260 --> 00:00:19,510 ‫First off one is a microphone. 4 00:00:19,510 --> 00:00:24,230 ‫Everyone is probably familiar with what a microphone is has one used one or heard one. 5 00:00:24,380 --> 00:00:28,040 ‫And actually that's what we're using right now to talk to you guys and microphone. 6 00:00:28,300 --> 00:00:31,130 ‫But what does a microphone actually physically do. 7 00:00:31,180 --> 00:00:37,660 ‫We know we talk to a microphone and it makes it broadcasts our voice over speakers correct. 8 00:00:37,870 --> 00:00:44,680 ‫Well not necessarily the microphone is simply a sound transducer and all that means is a microphone 9 00:00:44,680 --> 00:00:49,560 ‫takes and sound waves such as your voice and turn them into voltages. 10 00:00:49,810 --> 00:00:55,060 ‫And so the output from a microphone is going to be a signal that changes really fast just depending 11 00:00:55,060 --> 00:00:57,950 ‫upon the pitch or the volume of your voice. 12 00:00:59,250 --> 00:01:02,660 ‫So how do we interface a microphone with an FPGA. 13 00:01:02,910 --> 00:01:08,760 ‫Well just like a temperature sensor we need to have an analog to digital conversion that takes place 14 00:01:09,090 --> 00:01:12,650 ‫because the output of a microphone is an analog signal. 15 00:01:12,760 --> 00:01:21,330 ‫And so you have to either have an external ADC chip for an internal ADC if your FPGA supports that and 16 00:01:21,690 --> 00:01:26,640 ‫when the microphone puts an output the louder your voice the higher the output will be. 17 00:01:26,700 --> 00:01:31,690 ‫And depending on the frequency of your voice will determine the frequency of that signal. 18 00:01:31,950 --> 00:01:38,490 ‫And by taking samples of the microphone at regular intervals we can gain an idea of what that actual 19 00:01:38,580 --> 00:01:42,250 ‫signal looks like from a digital perspective. 20 00:01:43,190 --> 00:01:45,830 ‫Some will throw a little digital signal processing theory at you. 21 00:01:45,850 --> 00:01:53,400 ‫Here we have the Nyquist theory which is also known as the sampling theorem and it states that you must 22 00:01:53,400 --> 00:01:59,590 ‫sample at a minimum of twice the rate of a signal you're trying to accurately sample. 23 00:01:59,760 --> 00:02:01,990 ‫So what does that mean for a microphone. 24 00:02:02,100 --> 00:02:09,660 ‫We're assuming that we're trying to capture the human voice and humans can hear typically in the range 25 00:02:09,660 --> 00:02:12,750 ‫of 20 hertz to 20 kilohertz. 26 00:02:12,750 --> 00:02:19,560 ‫So to accurately sample a microphone and to be able to recreate that signal we need a sample at a rate 27 00:02:19,620 --> 00:02:23,830 ‫of twice 20 kilohertz which would be 40 kilohertz. 28 00:02:24,060 --> 00:02:30,030 ‫And if you notice when you use any type of software with like audio recording typically will have a 29 00:02:30,030 --> 00:02:36,050 ‫rate of 44 kilohertz and that's because you want to be at least minimum of twice the sampling rate. 30 00:02:36,150 --> 00:02:41,480 ‫So they'll sometimes say OK 40 is the minimum we'll go up to 44 just to be sure. 31 00:02:41,490 --> 00:02:47,860 ‫And there's a few people who can say they can hear higher than 20 kilohertz. 32 00:02:47,870 --> 00:02:55,370 ‫Now if we under sample the audio let's say we don't reach the full 40 kilohertz we sample at 20 kilohertz 33 00:02:55,370 --> 00:02:57,610 ‫or 30 kilohertz per se. 34 00:02:58,080 --> 00:03:03,190 ‫If we do this we won't be able to actually accurately recreate that signal. 35 00:03:03,260 --> 00:03:07,790 ‫We need to know what the highest frequency of the signal we are trying to capture is so that we can 36 00:03:07,790 --> 00:03:12,420 ‫accurately recreate that signal and get an understanding of what's going on. 37 00:03:12,820 --> 00:03:18,110 ‫If we're if we're under sampling and we think that something else is that is occurring and we're not 38 00:03:18,110 --> 00:03:19,270 ‫getting the full picture. 39 00:03:19,280 --> 00:03:24,170 ‫This is called aliasing and that's an digital signal processing. 40 00:03:24,170 --> 00:03:29,620 ‫They'll use the term aliasing now over sampled audio. 41 00:03:29,680 --> 00:03:33,700 ‫You cannot over sample audio where it hurts you per se. 42 00:03:33,940 --> 00:03:41,530 ‫If we say we need to have a signal sampled at 44 kilohertz we could sample it at 400 kilo Hertz 500 43 00:03:41,530 --> 00:03:43,190 ‫kilohertz 600 kilohertz. 44 00:03:43,360 --> 00:03:47,200 ‫It's not going to hurt you or you're just going to throw more samples on that signal and you actually 45 00:03:47,200 --> 00:03:50,580 ‫get a better idea of what's actually occurring. 46 00:03:51,000 --> 00:03:55,450 ‫And some people when you're working with audio signals they'll claim that when you're sampling at a 47 00:03:55,450 --> 00:04:00,150 ‫much higher rate it makes the audio sound much more pure or clear. 48 00:04:00,160 --> 00:04:05,430 ‫However to me at 44 kilohertz I can't tell the difference of that sample are higher. 49 00:04:05,470 --> 00:04:08,650 ‫Now what will happen if you over sample audio. 50 00:04:08,650 --> 00:04:11,650 ‫It's going to cost you in terms of processing time. 51 00:04:11,740 --> 00:04:20,950 ‫If you had only a much faster DSP microprocessor or a PDA in his course or with PJ's to run faster means 52 00:04:20,950 --> 00:04:26,260 ‫we're going to utilize more power or Also more resources so we can do more things in parallel. 53 00:04:26,290 --> 00:04:29,140 ‫So you want to sample high enough to recreate the signal. 54 00:04:29,260 --> 00:04:35,180 ‫What you don't want to oversample so much that it costs you in terms of resources and processing time. 55 00:04:36,400 --> 00:04:39,160 ‫Now that you know how to interface your FPGA with the microphone. 56 00:04:39,160 --> 00:04:40,660 ‫Let's get started on the assignment.